Re: [PATCH v2 2/3] xen/hvm: introduce a fpu_uninitialised field to the CPU save record

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 18/11/15 11:04, Jan Beulich wrote:
>>>> On 18.11.15 at 12:00, <andrew.cooper3@citrix.com> wrote:
>> As for the problem at hand, I don't see what was wrong with v1.
>>
>> Fundamentally, we have three different variations of the same structure;
>> two of which require special compat handling.  Pretending otherwise is
>> just silly.
> And if we gain a few more additions, we'll end up with half a dozen
> slightly different structure declarations in the public interface? If
> you consider the first step (adding tsc_aux) a mistake, let's not
> repeat the same mistake again (even if right now it might _seem_
> to be reasonable to some of us).

As part of cpuid handling fixes, I will be removing all of this from the
hypervisor

The result doesn't need to live very long; I would recommend the route
which has more-obvious-correct code.  I am not fussed as to which route
this ends up being.

~Andrew